Morphological analysis and filler distribution
First, the structures of the used carbon fillers, i.e., GNPs and carbon nanotubes, were investigated. Their SEM images are presented in Fig. 1a-h. The average lateral size of GNPs was confirmed to be 25 µm. Interestingly, in the case of MWCNTs, the A and B types have more homogeneous outer diameter distributions than the C type, for which more nanotubes with larger diameters are noted. Typical SEM images of the composite samples THV/GNP, THV/MWCNTs, and THV/GNPs/MWCNTs are presented in Fig. 1e-h. In the microscale SEM images shown in Fig. 1e-f, the filler distributions appear to be uniform for both composites. However, pictures taken at lower SEM magnification (Fig. 1g-h), supported by the polarizing-interference microscope image (Fig. 1i), show a different and interesting view. The carbon fillers agglomerate and orient in the polymer matrix, creating long, macroscopic, parallel paths. The paths are as long as over 1 mm (Fig. S1). This effect could be related to the relatively high melt flow index (265°C/5 kg) of the used THV matrix, which equals 20 g/10 min (measured by the ASTM D1238 method). During the hot pressing process, the polymer quickly melts and starts to flow, which promotes orientation of the fillers. But the flow may have also just a supporting effect. Important is the structural incompatibility of the polymer matrix and fillers. The C-F bond is strongly polarized and each such bond in the polymer chain is a dipole. Hence, THV is a polar polymer and carbon fillers having only C-C connections are non-polar. This difference in polarity leads to a strong phase separation, with a tendency to agglomeration within the phases. The high melt index may only make it easier. The orientation may be also related to the flow direction during pressing.
Continuing structural analysis, Fig. 2a presents the typical Raman spectrum of the THV/GNP composite. The following graphene-originated bands are easily observed: the D mode at 1315 cm− 1 (breathing mode A1g), G mode at 1602 cm− 1 (stretching mode E2g), 2D mode at 2618 cm− 1 (overtone of the D mode) and modes related to the THV matrix at 300, 375, 669, 716, 822, 897, 1054, 1436 and 2996 cm− 1. Because the THV terpolymer includes the monomeric units tetrafluoroethylene (TFE, C2F4), vinylidene fluoride (VDF, CH2CF2) and hexafluoropropylene (HEP, C3F6), its spectrum is expected to appear as a superposition of the spectra of PTFE and PVDF with additional bands originating from the boundary regions between the different monomers, as suggested by E. D. Emmons et al.. Based on this, the individual THV modes were marked as follows: 300 cm− 1 - γt(CF2) of PTFE; 375 cm− 1 – δ(CF2) of PTFE; 669 cm− 1 and 716 cm− 1 - νs(CF2) of PTFE; 822 cm− 1 - γr(CH2) of PVDF; 897 cm− 1 - νs(CF2) of PVDF; 1054 cm− 1 - να(CC) of PVDF; 1436 cm− 1 - δ(CH2) of PVDF; and 2996 cm− 1 - νs(CH2) of PVDF.
In the next step, to more precisely study the filler distribution in the THV matrix, Raman mapping was performed. Spatial analysis of the intensity ratio of the most prominent Raman modes for the polymer and carbon filler gives information about the homogeneity of the composite sample. Figure 2b,c presents the Raman maps registered for the intensity ratio of the THV band at 822 cm− 1 (named ITHV) and graphene 2D band at 2618 cm− 1 (named I2D). To simplify the reading of the map, the intensity ratios were normalized, wherein a dark green color shows graphene domination and a light green color shows polymer domination. Large dark green zones are visible on the map, indicating orientation of the filler in the polymer matrix, which confirms the results obtained by SEM and polarizing-interference microscopy.
XRD analysis
Subsequently, XRD was carried out to reveal any carbon-induced structural changes that THV may have undergone during the mixing process. One must note that to avoid modifying the materials further, since they are very soft and pliable in nature, the materials were not pulverized before the examination. The only exception was the reference THV sample that was in the form of a very coarse powder, which was very difficult to pulverize further even after pouring liquid nitrogen on it. Because the carbon-modified samples and the reference sample were used in the form of bulk or coarse powder, respectively, the results must be treated only qualitatively. Additionally, note that only the area within tens of micrometers from the sample surface provides the diffraction signal, so in the case of macroscopic long paths created in the polymer matrix composite, the relative intensities of peaks belonging to different phases may be very different from the expected intensities based on the quantity of the additive. Based on the abovementioned factors, further comparative analysis was not performed, as there were no exact descriptions of the nature of composite samples measured through XRD in the literature.
The results are summarized in Fig. 3. First, the reference spectra are shown in Fig. 3a. The THV spectrum shows very low crystallinity of the reference material, with only 3 very broad and very low-intensity peaks, suggesting a close-to-amorphous nature of the sample. The GNP spectrum reveals a typical graphitic structure with the P63/mmc(194) space group. On the normalized graph scale, only the peak corresponding to (002) is visible; however, closer magnification reveals a perfect fit with the remaining peaks in both position and relative intensity. This suggests that there is a significant amount of graphite in the GNPs. In the case of the XRD patterns obtained for the nanotubes, they, as expected, also show a peak corresponding to the (002) graphitic plane but a weak (101) peak.
Analyzing the composite samples shown in Fig. 3b-c, the THV patterns also exhibit only a close-to-amorphous character. The carbon fillers may be assumed to not promote or limit crystallite formation in the polymer matrix. In addition, no significant shifts are observed. Increasing the filler loading results in a stronger filler signal, which is clearly seen in Fig. 3b and 3c. The only exception to the rule is the very low intensity of the graphite peak in the case of the C1/G1 sample (Fig. 3c), which may be caused by reflection arising from the MWCNTs. In particular, the GNPs “stick” to the MWCNTs when used simultaneously, as shown in Fig. S1. The same observation was made for samples A1-C1, as shown in Fig. S2 obtained with a Cu Kα anode lamp, where weak MWCNT-originated reflections are visible. These observations are consistent with 33 and their observations of PVDF/GNP XRD patterns. This result shows that the fillers may be freely used to modify the matrix without any structural changes being introduced into the system.
Thermal analysis: thermogravimetry and DSC
TGA curves for the studied samples are presented in Fig. 4a-c, while DTG (temperature derivative of TGA curves) curves are presented in Fig. 4d-f. The characteristic temperatures are summarized in Table 3. In the reference sample (without any additives), one can observe the decomposition process of the pure polymer starting ca. 365°C and occurring at a maximum rate ca. 462°C (column Td_THV in Table 3). The Td_THV temperature was calculated by taking the DTG peak maximum into account.
For composites with GNPs only, the situation is slightly different (Fig. 4a and 4d). The Td_THV temperature varies from 440°C to 454°C among the different samples and increases with the GNP concentration. At higher temperatures, a second DTG peak appears. Its intensity increases with the GNP concentration; therefore, it may be assumed to be related to GNP decomposition. It is labeled Td_GNP in Table 3. Peak maxima are present in the ca. 489–498°C temperature range. When the temperature is increased even higher, another DTG peak appears. It is visible only for samples with higher GNP content (not less than 5%). The peak maxima are in the 731–786°C temperature range. This peak is labeled Td_GNP2 in Table 3 because it was similarly observed only for GNP composites. In the case of both Td_GNP2 peaks, there is no clear relation between the GNP content and temperature.
According to , graphene decomposition occurs in the temperature range of 400–800°C, which nicely fits with both obtained GNP decomposition peak temperatures. In Fig. 4a and 4d, two mass losses are observed. Distinguishing between polymer mass loss and graphene mass loss is, however, impossible, as the peaks related to these processes overlap. Therefore, the relative polymer and graphene mass loss cannot be evaluated. In all cases, both components were fully decomposed at a maximum temperature of 900°C.
TGA and DTG curves for samples A1-C1 (1% loading of various diameter MWCNTs) are presented in Fig. 4b and 4e, respectively. Three peaks can be observed on the DTG curves. The highest one is related to THV decomposition. This process starts ca. 370°C (similar to GNP samples), and the maximum decomposition rate occurs in the 452–454°C range. The second peak maxima are present in the 483–493°C range, and the third maxima are present in the 649–700°C range. The intensities of these peaks are similar for every sample, suggesting that they are related to MWCNT decomposition. Therefore, they are labeled Td_MWCNT and Td_MWCNT2 in Table 3. There is no clear relationship between the nanotube diameter and peak temperature.
For samples C1/G containing both GNPs and nanotubes, the thermal behavior is similar to that of the previous composites (Fig. 4c and 4f). The decomposition process of the polymer starts ca. 350°C for all samples (the lowest among the entire series). The Td_THV temperatures were evaluated to be in the 443–468°C temperature range. For all samples, there are two other DTG peaks. The intensity of the lower temperature peak is related to the GNP amount. The peak maximum temperatures (Td_GNP) are in the 491–498°C range (Table 3). Decomposition of the nanotubes occurs at similar temperatures; therefore, these two processes cannot be distinguished. The maxima of high-temperature peaks are in the 648–670°C temperature range, similar to peaks in the A1-C1 samples. Therefore, they may be ascribed to MWCNT decomposition. These temperatures are labeled Td_MWCNT2 for correspondence to the A1-C1 results. Finally, at the highest temperatures (804–821°C), a fourth DTG peak may be observed only for samples with GNP amounts not less than 5%, which is in accordance with the results for the G samples.
Table 3
Summary of TGA results of THV-based nanocomposites. Td_THV, Td_GNP, and Td_MWCNT represent the THV, GNP and MWCNT maximum decomposition rate temperatures.
Sample
name
|
Filler type
|
Filler
concentration
wt%
|
Td_THV
[oC]
|
Td_GNP
[oC]
|
Td_GNP2
[oC]
|
Td_MWCNT
[oC]
|
Td_MWCNT2 [oC]
|
ref
|
-
|
0
|
462.4
|
-
|
-
|
-
|
-
|
G1
|
GNP
|
1
|
453.2
|
-
|
-
|
-
|
-
|
G2
|
GNP
|
2
|
439.8
|
491.8
|
-
|
-
|
-
|
G5
|
GNP
|
5
|
450.2
|
497.4
|
752.6
|
-
|
-
|
G7.5
|
GNP
|
7.5
|
448.3
|
498
|
777.9
|
-
|
-
|
G10
|
GNP
|
10
|
442.4
|
497.1
|
785.6
|
-
|
-
|
G12.5
|
GNP
|
12.5
|
454.4
|
497.4
|
731.2
|
-
|
-
|
G15
|
GNP
|
15
|
440.3
|
488.9
|
762.4
|
-
|
-
|
A1
|
A-type MWCNT
|
1
|
463
|
-
|
-
|
492.5
|
648.7
|
C1
|
C-type
MWCNT
|
1
|
452.4
|
-
|
-
|
491.5
|
699.6
|
B1
|
B-type MWCNT
|
1
|
453.5
|
-
|
-
|
483.1
|
662.8
|
C1/G1
|
C/GNP
|
1/1
|
456.2
|
497.1
|
-
|
647.5
|
-
|
C1/G2
|
C/GNP
|
1/2
|
442.9
|
490.6
|
-
|
669.7
|
-
|
C1/G5
|
C/GNP
|
1/5
|
451.3
|
495.3
|
805.8
|
646.3
|
-
|
C1/G7.5
|
C/GNP
|
1/7.5
|
466.3
|
497.8
|
821
|
650.9
|
-
|
C1/G9
|
C/GNP
|
1/9
|
467.6
|
493.4
|
804.4
|
663.5
|
-
|
Figure 5 shows comprehensive DSC results for all the THV-based composites obtained in three cycles to reveal any filler-induced changes in the composite structure. Starting once again from the DTG curves obtained for GNPs show 3 peaks indicating local maxima of the degradation rate (Tmax): 306.8°C, 438.0°C and 738.2°C (Fig. S3). According to Farivar et al., the first peak occurs in the region of decomposition of oxygen-containing surface groups. The second peak could be related to the combustion of highly defective carbon layers, although it is shifted toward lower temperatures compared to the literature. The highest temperature peak refers to sp2 carbon layer decomposition. Interestingly, from the very beginning of the test, a small mass loss is observed, which suggests small-molecule compound evaporation.
In DSC, in the first heating run (Fig. S4), a distinct endothermic peak at approximately 0°C is observed, which is followed by a wide curve indicating another endothermic phenomenon. This could be ascribed to water melting and subsequent evaporation. Assuming that the melting enthalpy of ice (in bulk) is 334 J/g, the amount of evaporated water could be estimated as ca. 0.9 wt%, which is much less than that seen in TGA. Moreover, the effects of water phase transitions can also be seen in the subsequent DSC runs: even after heating to 240°C, in the third cycle, a melting peak of a similar enthalpy appears. This suggests that some of the adsorbed water must be strongly bound in the filler pores. If so, then the amount of evaporated water should not be calculated in the above way since the melting enthalpy takes a different value, and on the surface of the pores, a significant amount of water does not freeze25. However, by using the freezing point depression method, the average pore size can be roughly estimated. In the cooling run, the freezing point appears at 41.4°C. According to the equation given by Sun and Scherer, the average pore radius is approximately 2 nm. However, this value should be treated with caution since the equation assumes a certain pore shape, which we know very little about in our case.
In the case of MWCNTs (all types) (Fig. S4), no transitions were detected in the studied temperature range, particularly no signs of adsorbed water.
Next, the high-temperature behavior of THV and its composites was studied. The pure polymer underwent a single-step degradation with no measurable residue. The maximum degradation rate was recorded at 491.2°C, which reveals the exceptionally high thermal stability of THV. The addition of 1% A-type CNTs does not affect this value, but the same amount of GNPs decreases Tmax to 480.5°C. The value remains similar for a much higher GNP content, such as 12.5% (additional DSC curve for this concentration, see: Fig. S6) (484.3°C). Moreover, for the composites with GNPs, the residue remains at approximately 6% even at temperatures reaching almost 1000°C. This indicates that the substances produced in THV-GNP composite degradation must react and form highly stable products. The reaction could be the reason for the Tmax decrease. Another possible cause for the thermal stability reduction could be a “chimney” effect: filler agglomerates could contain free spaces, which could act as a quick route for the escape of volatile decomposition products, accelerating the weight loss. No signs of a barrier effect or the influence of the high thermal conductivity and heat capacity of the filler were observed.
The DSC analysis of pure THV revealed a glass transition at approximately 3°C (good compatibility with producer data) (Fig. 5a-c, 5g-i). This means that at room temperature, the material is well above Tg, which should contribute to its elasticity. A pronounced melting peak was observed at 118°C in both heating runs (while producer declares 115°C). According to the literature, in THV-type terpolymers, the VDF and TFE monomers are capable of crystallization, but both PVDF and PTFE melt at higher temperatures (158–197°C and 327, respectively). The lowering of the melting temperature could be caused by randomness of the unit sequences in the polymer chains (THV is a statistical terpolymer produced via radical polymerization). HEP units hinder the crystallization of such terpolymers28, so their presence could result in a Tm decrease. Such a change is beneficial in terms of composite processing. In the first cycle, a small endothermic peak can also be observed, which disappears in the second heating run. This could be a result of melting of smaller or more defective crystallites formed during processing. In the cooling cycle (Fig. 5b), crystallization occurs at 88°C.
The addition of MWCNTs of any type results in a lower melting enthalpy in the first run, which indicates that crystallization is hindered. The curve shapes in both heating runs are relatively similar.
The glass transition of the matrix occurs in a temperature range similar to that of the melting of water introduced with the GNPs, making assessment of the presumptive changes in the composite Tg impossible (Fig. 5a, 5c). As a test, a single sample (C1/G1) was dried at 40°C in vacuum and examined again (Fig. S4). The obtained Tg was very close to the value for pure THV (3°C), and Tm and Tc were comparable to the values determined before drying (117°C and 87°C, respectively, Fig. 5d, 5f). This indicates that none of the fillers influences the polymer chain mobility enough to affect the characteristic temperatures. The reason for this could lie in the filler distribution, which has been discussed on the basis of microscopic observations.
In general, melting of the composite matrix occurs at similar temperatures (116–118°C in the first heating cycle and 116–117°C in the second heating cycle), and its filler content dependence does not follow any obvious trend. In all samples, the Tm in the 2nd heating run is slightly lower, probably because the crystallites formed during relatively quick cooling are smaller and/or more defective. Nevertheless, the melting enthalpy remains similar in both runs, which shows how easily THV crystallizes. This can be important information in the future analysis of mechanical properties as a function of temperature.
Total EMI SE and electrical conductivity
In this section, both the EMI SE and electrical conductivity will be analyzed, as they are dependent on each other. Figure 6a provides deeper insight into the GNP-induced total EMI SE for different filler loadings. A significant result is that a saturation effect is observed, and a value of 23 dB at 5 GHz for 10 wt% GNP loading is registered as the maximum value obtained for this series. For the higher concentrations, i.e., 12.5 wt% GNP loading, an ~ 12% decrease is observed, and for 15 wt% GNP loading, an ~ 18% decline is observed compared to the best sample. Such behavior is rather unique, and usually, no saturation effect is observed, even for very high GNP concentrations or even MWCNTs. To our knowledge, only Qin et al. reported an EMI shielding saturation effect for Ag nanowire/PEDOT:PSS composites, but as a function of high-power microwaves, so it does not apply in our case.
Next, to investigate the synergistic effect of MWCNTs – GNPs in the polymer matrix, preliminary selection of the MWCNT diameter and length was executed by studying the EMI SE obtained for 1 wt% loading of different types of MWCNTs (see details in Table 1), as shown in Fig. 6b. Surprisingly, the size of the MWCNTs incorporated in the polymer matrix influences the total EMI SE, and the best effectiveness is observed for the C-type nanotubes, SETOT = 7.4 dB at 5 GHz, which is 75% and 32% greater than that for the A-type and B-type nanotubes, respectively. Similar observations were noted by Hu et al., who investigated buckypapers made of MWCNTs of different diameters and lengths: Ø = 10–20 nm and l = 10–30 µm; Ø = 8–15 nm and l = 1–100 µm; and Ø = 20–30 (15) nm and l = > 10 (3) µm, dispersed in water or ethanol. The greatest EMI SE was noted for the largest Ø = 20–30 (15) nm MWCNT, and this observation is consistent with ours. Another aspect that should be accounted for when analyzing the effect of CNTs on the EMI SE is their large aspect ratio, as mentioned by Wang et al. and Huang et al.. Therefore, to investigate synergistic effects, MWCNTs with Ø = 10–20 nm, l = 10–30 µm and a specific surface area of 233 m2/g were chosen (C-type nanotubes).
The synergistic effect in terms of EMI SE was investigated for composites of MWCNTs with a constant 1 wt% loading and GNPs with various loadings in the THV matrix (Fig. 6c). In contrast to the GNP-only series, no saturation effect is observed here. Moreover, a surprisingly low total SE is noted for the highest 9 wt% GNP loading, reaching 17 dB, which is a 23% lower value than for the GNP-only series. This suggests that there is no synergistic effect between the MWCNTs and GNPs, in contrast to, for example, acrylonitrile butadiene styrene (ABS)-based composites. Sharma et al. performed a similar investigation but based on ABS. Synergistic effects were revealed in the ABS-based composite when adding a 1 wt% MWCNT and GNP filler to the polymer matrix, reaching over 30 dB at 8 GHz for a 10 wt% GNP concentration. Simultaneously, SETOT = 23 dB at 8 GHz was achieved for the same GNP-only concentration. The higher values obtained in ABS-based systems may be directly due to differences in the polymer chains: ABS chains contain hydrocarbons on the end, while THV ends with trifluoromethyl and fluoride. The hydrocarbons may interact with both GNPs and MWCNTs more easily than with fluoride-based compounds. Moreover, a great influence of the synergistic effects may also be related to percolation path formation. In addition, synergistic effects between GNPs and MWCNTs in terms of EMI SE were found by Joseph et al., who obtained SETOT = 43 dB in the X band for a composition of 10 wt% of both GNPs and MWCNTs, and these effects were also reported in other publications. Regarding the particular behavior of the THV/GNP/MWCNT nanocomposite, a deeper analysis of the EMI shielding mechanism is needed.
The EMI SE values are for 1 mm thick samples.
Mechanism of the EMI SE in THV/GNP/MWCNT nanocomposites
Generally, the total SE is governed by three main mechanisms: reflection, absorption, and multiple internal reflection. Each of them originates in different physical phenomena, and reflection and multiple internal reflection (MIR) arise from impedance mismatch of the material and wave41. Figure 6d-f shows the reflection EMI SE obtained for all series considered in this article. For the majority of polymer nanocomposites, the reflection mechanism is small, and in the majority of cases, it is neglected, as its contribution to the total EMI SE is small. Nevertheless, to facilitate the analysis, Fig. 6g-i shows the comparison between the total EMI SE and reflection EMI SE at 5 GHz, which is the band used in wireless communication. The analysis does not account for MIR for several reasons: MIR may be neglected above 15 dB and when the sample thickness is smaller than the incident wavelength, or for higher frequencies, which applies within this work. The results shown in Fig. 6g-i clearly indicate the negligible contribution originating from the reflection mechanism. There is no value larger than SER = 0.42 dB obtained for 1 wt% MWCNTs of the C type; therefore, absorption may be assumed to be the main mechanism of the EMI SE in THV/GNP/MWCNT nanocomposites. Absorption is governed by the interaction of the electric and magnetic dipoles of the material with incident radiation. Moreover, it depends on the relative permeability, electrical conductivity, and frequency (\(A=8.686{t}_{n}\sqrt{\pi f\mu \sigma }\), where \({t}_{n}\) is the shielding material thickness, \(f\) is the frequency, \(\mu\) is the relative permeability, and \(\sigma\) is the electrical conductivity). The absorption mechanism also depends on the skin depth, which describes the depth at which the strength of the wave exponentially propagating through the conductive medium decreases to 1/e. The skin depth is
To more precisely determine the effects observed for the EMI SE, particularly for the THV/GNP series with a significant saturation phenomenon, the electrical conductivity for this series was calculated based on volume resistivity measurements, as shown in Fig. 7a,b. The character of the curve is similar to that for the EMI SE, from which a strong impact on the SETOT originating from the electrical conductivity may be assumed. However, to obtain a more explicit description of the electrical conductivity, the effective electrical conductivity was calculated, as shown in Fig. 7c. The obtained relationship agrees well with the classically used power law, which can be given for THV-based composites by the formula \(\sigma ={\sigma }_{0}{\left(\frac{\varphi -{\varphi }_{c}}{1-{\varphi }_{c}}\right)}^{t}\), where \(\varphi\) is the filler volume fraction relative to the total volume, \({\varphi }_{c}\) is the critical volume fraction, \({\sigma }_{0}\) is the conductivity of the filler, and \(t\) is the critical exponent. The formula works for \(\varphi\)> \({\varphi }_{c}\), and clearly, the percolation threshold can be roughly estimated to lie below 1 wt% GNP loading, which also agrees with the percolation threshold of other composites. Additionally, a wide and long plateau is observed above the percolation threshold, with no strong saturation effect. Magnifying the region above the percolation threshold, there is some fluctuation for the 10 wt% GNP loading, but saturation at this specific filler concentration cannot be stated. Therefore, further comprehensive analysis needs to be performed, especially in terms of magnetic permeability and probably also structural analysis, which is beyond the scope of this work.
Thermal conductivity and multifunctionality
Next, the thermal conductivity was measured, as shown in Fig. 8, combined with the electrical conductivity and total EMI SE at 5 GHz. In contrast to the electrical conductivity discussed above, the thermal conductivity does not show a significant and strong percolation threshold. In the case of the thermal conductivity, percolation network formation depends greatly on the shape of the filler, filler packing factor, filler aspect ratio, gaps between fillers, phonon scattering, and thermal resistance and is usually obtained for higher filler concentrations. Therefore, the explicit description of the thermal transport in polymer composites is rather complex. Moreover, the thermal conductivity threshold and percolation models remain an open issue; there are works describing the relevance of using percolation models in the case of polymer composites, such as Kargar et al., who reported thermal percolation threshold achievement for GNP/epoxy composites at a high 35 wt% GNP loading according to Lewis–Nielsen models, or Shtein M. et al., who observed that the percolation threshold correlated with a power law for GNP/hexagonal boron nitride (hBN)/epoxy composites53. However, there are also works that have described results that follow a linear dependency, such as those by Shahil, K. M. F. & Balandin, A. A. and Shenogina et al., who investigated carbon nanotubes and graphene-based polymer composites. The GNP/THV composites also follow a linear dependency, with no signs of a percolation threshold. There may be several reasons for this. First, the results shown here are for relatively low weight GNP concentrations, and the maximum GNP loading is 15 wt%, while Kargar F. et al. observed a percolation threshold of 35 wt% GNPs54. Next, THV is a different type of polymer matrix – thermoplastic with ρ = 1.97 g/cm3 – while those reported in the literature are usually thermosets of lower density epoxy resin: ρ = 1.2–1.3 g/cm3. The processing procedure may also affect the final thermal conductivity performance. Of course, one needs to remember the issues that may influence thermal conductivity path formation described in the beginning of this paragraph. Despite the lack of a percolation threshold, the maximum thermal conductivity of THV/GNP is κ = 1.65 W/mK, providing over 800% enhancement of this feature compared to the bare matrix (κ = 0.202 W/mK). Again, slight fluctuations are also observed for the 10 wt% GNP loading, but this issue will be clarified in the future.
Finally, multifunctionality, as the main aim of this work, will be briefly discussed. The idea of designing multifunctional materials is to combine several functions in one device/system and simultaneously decrease the weight of the whole device/system. This issue is critical in airborne and space industries, which struggle with overweight noneffective EMI shielding and overheating of avionics systems. Designing materials with excellent thermal conductivity, excellent EMI SE, and low mass may solve these problems and additionally provide longer flights, shorter takeoff distances, shorter landing distances, more economical flights, and lower CO2 emissions into the atmosphere. Fluoropolymers are attractive for airborne and space applications due to their enhanced basic performance, and the THV/GNP composites provide good thermal conductivity and EMI SE, having a much lower electrical conductivity than contemporary metals, as shown in Fig. 8, and being more resistant to harsh environmental conditions. Recall that the THV/GNP composites provide the maximum SETOT = 23 dB at 5 GHz for a 1 mm sample thickness, with κ = 1.65 W/mK and σ = 1.49 S/cm, for 15 wt% GNP loading. In our latest publication concerning PVDF/GNP composites, 15 wt% GNP loading provided SETOT = 33.29 dB (at 5 GHz), κ = 2.47 W/mK, and σ = 3.02 S/cm. In contrast, Barani et al. presented the EMI SE and thermal conductivity for GNP/epoxy composites, and for sample above 15 wt% (9.5 vol%), they obtained SETOT < 30 dB at 10.2 dB (SETOT ~ 15 dB for THV/15 wt% GNP at 10.2 GHz) and thermal conductivity κ ~ 2.5 W/mK. There are also other works on the multifunctionality of polymer composites such as polylactic acid (PLA)/GNP, polydimethylsiloxane (PDMS) GNP foam, and others, and the diversity obtained is actually very wide. There are results much better than those presented here but also lower than those shown here. One needs to remember that the THV matrix has excellent flexibility, much lower processing temperatures than other fluoropolymers, and much better basic performance in terms of flammability or chemical resistance, which is a great advantage, especially compared to PLA, epoxy, etc. Therefore, the results obtained for THV/GNP composites make them attractive for future advanced applications, especially in EMI shielding and thermal management.